ML20056H575
| ML20056H575 | |
| Person / Time | |
|---|---|
| Site: | Comanche Peak  |
| Issue date: | 09/03/1993 |
| From: | William Cahill TEXAS UTILITIES ELECTRIC CO. (TU ELECTRIC) |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| TXX-93321, NUDOCS 9309100104 | |
| Download: ML20056H575 (24) | |
Text
l L" =~~" Log # TXX-93321 File # 10010 r
=___
TUELECTRIC September 3, 1993 William J. Cahill, Jr.
Geemp Vice President U. S. Nuclear Regulatory Commission Attn: Document Control Room Washington, DC 20555
SUBJECT:
COMANCHE PEAK STEAM ELECTRIC STATION (CPSES)
DOCKET NOS. 50-445 AND 50-446 REQUEST FOR ADDITIONAL INFORMATION CONCERNING CPSES TOPICAL REPORT RXE-91-005, " REACTOR CORE RESPONSE TO STEAM LINE BREAK EVENTS" REF: TU Electric Letter logged TXX-91206 dated May 31, 1991, from W. J. Cahill Jr., to the NRC Gentlemen:
By letter dated July 7, 1993, the NRC staff requested additional information concerning the CPSES Topical Report RXE-91-005 submitted in the referenced TU Electric letter. A response was requested within 30 days of receipt of the NRC letter. Based on subsequent telephone conversations between the NRC staff and TU Electric personnel, the response submittal date was rescheduled. Enclosed is the requested information.
l Should you have any questions concerning this submittal, please contact Bob Dacko at (214) 812-8228.
l Sincerely, i
1 William J. Cahill, Jr.
By:
D. R. Woodlan Docket Licensing Manager BSD Enclosure c- Mr. J. L. Milhoan, Region IV ResidentInspectors,CPSES(2)
Mr. T. A. Bergman, NRR Mr. B. E. Holian, NRR i
j Dr. Heidi Komoriya l
International Technical Services 420 Lexington Ave O New York, NY 10170 9309100104 930903 5
D 0 DR ADOCK 0500 400 N. olive street LB. Ei Datias Texas 75201 ,
i ENCLOSURE TO TXX-93321 ,
(23 PAGES) s TU Electric Responses to the !
NRO's Request For AdditionalInformation Concerning i i
" Methodology for Reactor Core Response l to Steamline Break Events" l
RXE-91 -005 f Question 1: !
t identify the version of RETRAN-02 used for aralyses presented in the topical ,
report. Discuss how the reactivity coefficients / curves used in RETRAN are !
generated. l Response: i RETRAN-02 MOD 005 was used. ,
The main steam line break (MSLB) is a cooldown event and is analyzed at end ,
of cycle (EOC) when the moderator temperature coefficient is most negative.
For each reload core configuration,- a reactivity comparison between the ,
RETRAN-02 analysis and three dimensional cycle specific neutronics calculations (discussed in more detail in the response to Question 7) is used to verify that the RETRAN-02 analysis remains conservative. If, through the reactivity comparison, it is determined that the RETRAN-02 analysis is no l longer conservative, new reactivity coefficients are calculated using the ,
methodology described in Reference B. Three-dimensional, full core SIMULATE-3 calculations are performed for reactor core conditions representative of the l MSLB. The SIMULATE-3 three-dimensional model has all control rods in the i j core except for the most reactive control rod. Using this model, the core l average moderator density coefficient is determined over the expected density i range of the event. The core average Doppler only defect is determined for fuel i temperatures corresponding to the-expected power range. ' As stated in the l Topical Report [A), all reactivity coefficients have added conservatism which !
accounts for the uncertainties identified in Reference B. >
1 l
l
I Question 2:
Identify any changes made by TU Electric to the generic VIPRE model.
! Response:
1 Because both Question 2 and Question 10 relate to the VIPRE modeling methodology used in the Steam Line Break analysis, they are answered together. TU Electric does not have a generic VIPRE model. The VIPRE model used for the at-power steamline break analysis and described in detail in Reference C, a 1/8 core model, is presented to demonstrate TU Electric's basic VIPRE modeling methodology in accordance with the VIPRE SER requirements.
The VIPRE model used for the Steam Line Break at Hot Shutdown, hereinafter l
referred to as the full core model, retains that basic TU Electric VIPRE modeling ,
l methodology. There are, however, two differences between the VIPRE models used for the Steam Line Break at Power and at Hot Shutdown analyses. The first of these is the radial noding scheme and the second is the axial noding scheme. The details of the different noding schemes are described below along with the reasons for the changes.
The radial noding scheme used for the full core model is just an expansion of the radial noding used in the 1/8 core model for the at-power analysis. The radial noding was expanded to facilitate modeling of asymmetric core inlet conditions, but the level of detail around the hot subchannels was preserved.
For the at-power MSLB event, all four loops are equally affected by the steam line break; therefore, a 1/8 core model which takes advantage of the resulting core inlet symmetry to reduce the number of radial nodes is utilized. For the SLB at Hot Shutdown event, complete mixing of the four loop flows in the inlet ;
plenum is not always assumed. This leads to asymmetric temperature distributions at the core inlet. Although it would be possible to model this asymmetric temperature distribution with a 1/8 core model, the asymmetric conditions are more easily input by defining the radial nodalization to be consistent with the boundaries of the asymmetry.
2
Figure 2-1 shows the radial noding for the at-power MSLB VIPRE model and Figures 3.4-5 and 3.4-6 in the Topical Report [ A] show the radial noding for the SLB at Hot Shutdown model. In both models, the hot thimble channel and the ,
hot standard channel, are modeled such that two subchannels are modeled adjacent to them in any lateral direction. Thus, the detailed nodalization for the hot subchannels and their surroundings, as described in the VIPRE SER as necessary to adequately resolve the flow field in the vicinity of the hot channel is preserved in both models. In the 1/8 core model, the remainder of the hot assembly is modeled as one lumped channel whereas in the full core model, two additionallumped channels of 12 flow cells each are modeled.
Moving away from the hot assembly, each assembly surrounding the hot assembly is modeled as a separate lumped channelin the full core model. This was done to facilitate use of this model in future analyses of mixed cores in which the fuel types may vary around the hot assembly. In the 1/8 core rnodel, all eight of these assemblies are represented as a single lumped channel since all eight assemblies are identical with respect to fuel type, power, and core inlet conditions. The remainder of the core in both VIPRE models is represented by two lumped channels. For the full core model, these lumped channels are defined to model the remainder of the hot quadrant of the core and the remaining approximately 3/4 of the core, respectively. This providas the ability to conveniently model quarter-core radial asymmetries.
With respect to modeling asymmetric core conditions, core inlet asymmetries ;
are input into VIPRE by describing the flow and temperature of the coolant entering each modeled subchannel or channel. Thus, the use of the full core model simplifies the representation of conditions in which the flow and/or the temperature of the RCS are significantly different in one quadrant of the core i than the remainder of the core by allowing one value of flow and/or temperature to be specified for channels 1 through 30 and different values to be specified for channel 31 (refer to Figures 3.4-5 and 3.4-6 in the Topical Report [AJ). VIPRE then resolves the mixing of the flow as it progresses up the core based on the crossflow diversion and turbulent mixing parameters input by the user. Radial power asymmetries are modeled by specifying the radial '
power peaking factors for each of the fuel rods represented in the model.
3
t r
i
) !
) 1 Channe6 4. Hot Thimete Channel
)2 2 3 -
3 Channet 7. Hot Standard Channet i
)4 5 5
6 6
7.- l 8 9 10 11 1
)O ~
) O,,0 00 -
20@O0000 !
i.
h l Channess 1 13 1
\ !
I
~
le t 1
' i 115
' i i
e 1
...p@___q.___ : : i l
- i e
, l
$.
- I l
I e l l
.....__.r...y____r___
8 e , e .
' ' ' 8
, i 16 .i . .
._ _ ' _ _ . .:. . 4 . . _ _'. - - _ ;. _ _ 5 .
!@i
. _ .,. . . . .l . . . . ; . . .
i i !.
g i I
, e a 8 I e, , .
Figure 2-1 CPSES Vipre-01 Model Radial Noding 4
The axial noding differences between the two VIPRE models arose from a desire to provide a convenient transferral of the axial power profiles generated by the Reactor Physics models into VIPRE input. The Reactor Physics models produce axial power profiles in terms of average power factors in each six inch node of the active fuel region. The input of these axial power shapes into VIPRE is facilitated by designing equivalent axial nodalizations between the Physics models and the VIPRE models. For the at-power cases, a bounding axial power profile was used.
l To summarize, TU Electric applies a consistent VIPRE modeling methodology l for both the at-power SLB and zero-power SLB DNB analyses. Differences may l exist in the models used in the analysis, but in all cases, the models are l consistent with TU Electric's approved VIPRE modeling methodology. In the case of the VIPRE models presented in the Steam Line Break Topical report, different radial and axial nodalizations were used for the at-power and at-hot shutdown analyses. TU Electric has performed calculations covering a broad range of operating conditions (e.g., high power, low RCS flow, high core inlet temperature, and low pressure) to assess any differences in the MDNBR predictions of the two different VIPRE models. The results indicate that the predicted MDNBRs are within 1% or less of each other and the predicted axial locations of the MDNBRs are consistent with each other.
Question 3:
1 Identify the initial steam generator (SG) liquid mass level used in the analysis l
and discuss the impact of using different SG mass.
Response
The initial steam generator mass used in the at-power steam line break analyses was 104,450 lbm per steam generator at full power. Because the time periods of interest (near the time of reactor trip) occur well before the steam generator inventory is depleted, the at-power scenarios are insensitive to the initial steam generator inventory.
The initial steam generator liquid mass used in the HZP reference case was 173,500 lbm. This value represents 110% of the design HZP SG mass provided by the NSSS vendor. Through the parametric studies described on page 3-43 of Reference A, it was shown that the use of a higher than actual SG inventory is conservative for this event.
5
l Question 4: i Explain why a break upstream of a main steam isolation valve at-power was not analyzed. Similarly, explain why a single affected loop case was not ;
considered as part of at-power cases. t
Response
As shown in Figure 3.3-1 of Reference A, the break location fer at-power l steam flow increase events was upstream of the MSIVs; however, the actual l l
break location is not important for the at-power steamline breaks. Due to the I latent heat in the RCS which retards the cooldown and reduces the addition of l positive reactivity, the " return to power" aspects of the at-power MSLB !
transients are bounded by the zero power MSLB analyses. The at-power cases !
l are analyzed to demonstrate that the DNBR acceptance criterion is satisfied i
! near the time of reactor trip. For these scenarios, the MSIVs are open; hence, ,
a MSLB at any location has a nearly symmetric effect on the RCS response.
Because the reactor trip will occur either upon receipt of the same signals (e.g., 3 low compensated steamline pressure safety injection and main steam isolation f signals) or, more likely, on trip signals generated before the steamline isolation i signal, and, in all cases before the MSIV can be closed, there is no possibility !
of creating a significantly asymmetric response in the RCS.
l l
Question 5: l Discuss the implications of, and justify not splitting, the bypass region in a [
split-core model. :
1 Response: j The core bypass is $= 5% of the total RCS flow and represents fluid which is not available for core heat removal. The core bypass fluid removes no heat I from the core, and thus, has no effect on the reactivity response of the core.
The core bypass may actually promote interloop mixing; however, because the amount of interloop mixing is controlled by the analyst (see the Responses to Questions 6 and 12), the presence of the bypass region has no affect on the transient. Thus, it is considered acceptable to model a single bypass region.
i i
l i
I l i
( l
_ _ , ..-_.I
i Question 6:
Justify use of a nonconducting heat exchanger to achieve the " desired" core inlet fluid conditions. How are those fluid conditions determined?
Response
The nonconducting heat exchanger is used by the analyst to pre-set the amount of interloop mixing assumed for the upper and lower plena in the reactor vessel.
A RETRAN control system has been developed to control the amount of interloop mixing. When, due to mixing in the reactor vessel downcomer and lower plenum, the fluid entering the " faulted" quadrant of the core is comprised of 80% " faulted" fluid and 20% " intact" fluid, the RETRAN control system is used to calculate the resultant enthalpy, based upon a weighted average of the enthalpies of the faulted and intact loop fluids. A nonconducting heat ,
exchanger is then used in the lower plenum to add (subtract) the required !
amount of energy of the appropriate lower plenum volume, subject to the l constraint that the amount of energy added to the faulted lower plenum volume l must equal the amount of energy removed from the intact lower plenum volume. A similar system is used to control the amount of interloop mixing in t the upper plenum. ;
The " desired" core inlet fluid conditions (i.e., faulted and intact quadrants' inlet i enthalpies) are dependent variables which are functions of the assumed amount l of interloop mixing. The justification for the assumed amount of interloop mixing is provided in the response to Question 12.
I l
l l l
)
i r
i 7 i ;
l
~
Question 7:
Discuss in further detail why the RETRAN-02 point kinetics model is ;
conservative with respect to the three-dimensional neutronics calculations.
Response
Conservative core average reactivity coefficients are used by TU Electric in RETRAN-02 to predict reactor core conditions as a function of time after a MSLB in one reactor 1000, in contrast, the three-dimensional neutronics i calculation performed for the full reactor core considers non-uniform effects j associated with the specified MSLB state point conditions obtained from :
RETRAN-02. The neutronics calculation uses the methodology of Reference B.
Prior to the MSLB, the reactor is at end-of-cycle, hot zero power (HZP), with all control rods except the most reactive control rod inserted. The MSLB is i assumed to occur in the reactor loop adjacent to the core quadrant with the '
stuck rod, and the calculation is performed for RETRAN-02 statepoint conditions corresponding to the maximum power reached during the event. l The three-dimensional calculations consider non-uniform effects such as the ;
increased fuel and water temperatures in the vicinity of stuck rod and non-uniform core inlet temperature and flow conditions between the faulted and L intact coolant loops. Comparison of the maximum MSLB reactivity insertion i determined from the three dimensional neutronics calculations with that from '
the RETRAN-02 calculation confirms the conservatism of the RETRAN-02 calculation. A much smaller reactivity insertion is typically calculated with the j three-dimensional calculation than with the RETRAN-02 calculation. l l !
I t
l 6
l l
l 8 ,
i :
i i j
i
]
l Question 8: !
i j As required by the safety evaluation report on RETRAN-02, the boron transport ;
- model must be qualified before it can be used in licensing calculations. TU !
j Electric should justify why this has not been done or provide the qualification
- of the boron transport model.
Response
! The RETRAN-02 boron transport model is activated for the analysis of the MSLB event; however, due to the progression of the MSLB transient and because the input to the model is biased in a conservative manner, the degree of accuracy of the model has no effect on the conclusions of the analysis.
Therefore, there is no need to qualify the boron transport model. !
4 3
] Using the RETRAN control systems, the time at which boron is allowed to enter the core is delayed by conservative allowances for the times required to initiate !
safety injection, including diesel generator start times (if necessary), pump :
startup, pump suction realignment to the RWST, purge the safety injection [
piping of dilute fluid, and then the time to transport the boron from the injection point in the cold legs to the core. This total delay time is determined through I calculations external to the boron transport model. Only after this total delay ,
time has elapsed is the boron transport model activated. !
The steamline break return-to-power excursion is limited through Doppler '
feedback prior to the time of boron addition. For example, the reference case i
< discussed in the Topical Report attains a power plateau at 20.9% RTP '
approximately 100 seconds before boron is allowed to reach the core. The rate of change of the core power during this 100-second plateau is less than
+0.005% RTP/second. While at this plateau, a near-balance is achieved between the moderator temperature reactivity feedback and the sum of the ,
initial shutdown margin and the Doppler reactivity feedback (the continued addition of auxiliary feedwater to the steam generators prevents this plateau from being a true steady-state condition). The peak power actually occurs at the time when boron reaches the core; however, the difference between the minimum DNBR calculated at the time of peak power and at any other point on the plateau is insignificant. Thus, the timing of when boron actually enters the core has no significant effect on the conclusions of the analysis.
The boron transport model is used to demonstrate that the core returns to a subcritical condition; however, because quasi-static fluid and power conditions have already been attained, the exact timing of boron injection and the subsequent time-dependent boron concentration is unimportant to the conclusions of the analysis.
9
i i '
j Question 9:
. Discuss the method by which the axial power distributions used in the DNB f analysis are determined. Justify that the core conditions used for such determination are reflective of the conservative steam line break predictions.
The topica! report seems to imply a generic approach rather than a cycle j specific approach; explain and justify.
]
Response
The calculation discussed in Question 7 which provides the three-dimensional prediction of the reactor core characteristics for the MSLB includes core average, hot assembly, and hot pin axial power distributions for use in the DNB analysis. The axial power distributions for the DNB analysis are determined on ;
a cycle-specific basis. f' 1
1 4
l Question 10:
Describe in detail how the VIPRE 31-channel model accounts for asymmetric ,
core conditions. Explain thoroughly why the number of axial nodes was ;
reduced in the hot-zero power case from that used in the at-power cases.
I
Response
i See the Response to Question 2.
i 10 i
1 i
~
. l Question 11:
TU Electric stated that the SIMULATE-3 code is used to calculate the axial and radial power distributions based on the statepoint conditions from RETRAN-02 for use in the VIPRE-01 analysis. Discuss how the axial power distribution used in the RETRAN-02 analysis was determined and describe the iteration procedure between RETRAN and SIMULATE-3. Identify the locations and the time of maximum departure from nucleate boiling ratio (DNBR) for the cases analyzed.
Response
Because the RETRAN calculations are based on a point-kinetics model, it is necessary to ensure that the RETRAN point-kinetics model overpredicts the change in reactivity, thereby producing a conservative transient. The RETRAN reactivity check is performed using the SIMULATE-3 three-dimensional full core neutronics code with statepoints generated by RETRAN. A description of the iteration procedure between RETRAN and SIMULATE-3 is provided in the response to Question 1.
The second part of this question requests the locations and times of the Minimum DNBR (MDNBR) for the cases analyzed. For the at-power steamline break cases, the time of MDNBR, where applicable,is provided in the Sequence of Events for each of the cases presented (Tables 3.3-4 through 3.3-11). In all the at-power cases, the MDNBR occurred in the top half of the core and usually about 2/3 of the distance up the active fuel length.
For the Steam Line Break at Hot Shutdown cases, the time of MDNBR coincided with the time of the peak core heat flux. Figures 3.4-10, 3.4-15 through 3.4-26, 3.4-28 through 3.4-30, '3.4-32 through 3.4-54, 3.4-56 through 3.4-59, and 3.4-61 through 3.4-66 of the Topical Report [A] provide the core heat flux as a function of transient time from which the time of MDNBR can be inferred. For all the Hot Shutdown cases analyzed with the exception of the loss of offsite power case, the SIMULATE-3 calculations yielded a top-peaked axial power profile. In these cases, the axiallocation of the MDNBR was in the top half of the core. For the loss-of-offsite power case, SIMULATE-3 predicted a very bottom-skewed power profile. As a result, the point of MDNBR was below the bottom mixing vane grid. However, the analyses are valid because the W-3 correlation, which is used for the zero-power analyses, was developed based on closed channel DNB test data, i.e.,
DNB test data which did not rely on the benefit of mixing grids, in all DNB analyses, the axiallocation of the MDNBR is reviewed to ensure it falls within the applicability of the respective CHF correlation being used.
11
~
^
Question 12: l Justify the amount of inter-loop mixing assumed, j
Response
l The amount of interloop mixing assumed for the reference case (80%/20% mix l to the core quadrant with the stuc'K rod, and 73%/27% mix in the upper i plenum) is consistent with values reported by the NSSS vendor as used in the ;
current FSAR analyses. The adequacy of these assumptions was confirmed by >
TU Electric through a series of parametric studies described on pages 3-36 through 3-39 of the submittal. j The series of parametric studies presented in the Topica! Report are tabulated {
in Table 14 (provided in the Response to Question 14). Included in these
~
studies are various assumed amounts of mixing in the reactor vesselinlet and ,
outlet plena. From these studies, the following observation is made: -
In general, as the amount of mixing is increased, the temperature ,
of the intact quadrants reach lower values, and the minimum temperature in the faulted quadrant increases. Due to differences l in the relative contribution to the total reactivity from the intact and faulted quadrants as the temperatures change, an inflection point in the maximum power is evident. (This effect is also influenced by the 50/50 reactivity weighting between the faulted quadrant and the three intact quadrants.) However, as the amount of mixing is increased, the total vessel average temperature decreases and the minimum DNBR increases, even !
though the peak power may be slightly higher.
With respect to the relevant event acceptance criterion (minimum DNBR), the f extreme, no-mixing assumptions result in the most severe transient; however, !
as concluded on page 3-51 of the submittal, the minimum DNBR is relatively .
insensitive to pessimistic changes in the amount of interloop mixing assumed. l Therefore, it is concluded that the more reasonable mixing assumptions used for the reference case are adequate. ;
I i
12
i l
Question 13:
Discuss the reasons leading up to the turnover of the power peak prior to the ,
boron reaching the core. ,
i
Response
The core power excursion is influenced by the addition of positive reactivity l from the negative moderator temperature coefficient and by the addition of negative reactivity from the Doppler fuel temperature coefficient. The effect of the moderator temperature coefficient results in the loss of shutdown margin and subsequent criticality. The effect of the Doppler temperature coefficient '
limits the maximum power achieved as a result of the MSLB event. The l reduction in the rate of the steam generator depressurization minimizes the ;
continued addition of positive reactivity, which in turn results in a power l l plateau when the Doppler reactivity feedback matches the moderator reactivity feedback less the initial shutdown margin. As shown in Figure 13-1, the addition of baron occurs well after this plateau is attained. ,
i l
l l
13
REFERENCE SLB CASE REACTIVITY EDITS E R 5 g pq R H I ru B - Moderator C - Doppler o D - Boron _
O -
E - Total React.ivity Feedback E
O bo _ -
>S F-H -
> ~
T- R A ' R R r H
NL (N
O S
% $7 E
c c e c/
~
l i I I I I I I 70 40 80 120 160 200 240 280 320 TIME (SEC)
Figure 13-1 f.
i l
l Question 14:
Summarize results of parametric cases in a tabular form. Identify and discuss initial conditions (including the uncertainties) and transient assumptions resulting in the worst (DNBR) case and the highest return-to-power case. ,
Response
Presented in Table 14 is a summary of the sensitivity studies presented in l Reference A. The relevant event acceptance criterion is the Minimum DNBR, .
l which is compared to the design limit of 1.45. Although the peak power is I l provided for comparative purposes, the " faulted" and " intact" core inlet l temperatures also affect the calculated MDNBR.
The parametric studies summarized in Table 14 are provided either:
- 1) to demonstrate that the relevant acceptance criterion is insensitive to the i value used, within the range of the sensitivity study; or,
- 2) to demonstmte that the assumptions used in the reference case are '
conservativ-
! l l The ascumptions/ initial conditions used for the zero-power MSLB reference case described in Section 3.4.2.2 are intended to be used to demonstrate the acceptablity of a particular reload core configuration. The initial conditions used 'or the zero-power analysis are summarized in Table 3.4-2. Based on the t l sensitivity studies summarized in Table 14, and consistent with the current MSLB evaluation presented in the CPSES FSAR, no uncertainties on the initial l conditions of pressure, temperature, and pressurizer water level are used in the l reference case. Furthermore, through the parametric studies, it is concluded i
that the reference case provides a conservative wpresentation of the MSLB event.
l i
15 i
I
l Table 14 Summary of Zero Power MSLB Sensitivity Studies l -
! Figure Case Description Peak Faulted intact MDNBR
! No. Power T-in Tin I
3.4.x (% RTP) (* F) (' F) i 10 Reference Case 20.9 441 505 1.83 i 15 Moderator Density Defect + 10% 22.2 445 509 1.82 l 16 Moderator Density Defect - 10% 19.9 436 498 2.03 17 Doppler Defect + 10% 20.6 440 503 1.88 f
18 Doppler Defect - 10% 21.3 442 506 1.79 19 Nominal Boron Worth + 10% 20.9 441 505 1.84 l 20 Nomenal Baron Worth - 10% 20.9 441 505 1.84 1
21 1.3% Shutdown Margin 23.0 446 511 1.72 22 2.4% Shutdown f.iargin 16.9 425 483 2.41 l 23 Reactivity Weighting = 25% to Faulted Sector 18.8 433 494 2.02 i
24 No Re ictor Vesselinlet Mixing 22.5 414 520 1.78 l
25 50% Reactor Vesselinlet Mixing 20.9 462 483 2.17 26 100% Reactor Vessel inlet Mixing 21.2 472 472 2.13 l
28 No Reactor Vessel Outlet Mixing 21.0 429 513 1.80 29 50% Reactor Vessel Outlet Mixing 21.9 446 496 1.95 l 30 100% Reactor Vesse! Outlet M eng 22.7 449 490 2.06 32 No Main Feedwater nm; Reference AFW Distribution 20.7 442 506 1.87 I 33 Na Mar reedwater Flow; Equal AFW D.stnbution 20.9 442 503 1.86 34 No Main Feedwater Flow; No AFW Flow 19.1 445 508 1.90 35 Reference Main Feedv : U tiow; Equal AFW Flow 21.5 442 501 1.86 36 Faulted Loop Main Few '00% Nominal; Reference AFW Flow 20.7 442 506 1.87 l
37 Faulted Loop Main Feedwaise = 220% Nominal; Reference AFW Flow 20.7 442 506 1.87 38 Fau!ted Loop Main Feedwater = 250% Nominal: Reference AFW Flow 20.6 442 506 1.87 39 Pressurizer on Faulted Loop 21.0 441 504 1.88 ,
I 40 No SG Tute Bundle Height Reduc-ir = 20.9 441 505 1.84 i 41 Isoenthalpic Expansion Choking Model for Bre.sk Flow 20.9 441 505 1.84 42 initial Pressurizer Pressure + 30 psia 20.9 441 504 1.85 43 initial Pressurizer Pressure - 42 psia 21.0 441 505 1.84 44 initial Pressurizer Water Level- 5% span 20.7 440 504 1.98 16 I
d
45 initial Pressurizer Water Level + 5% span 21.2 442 506 1.93 ,
46 Narriinal Steam Generator Mass 21.0 441 504 1.84 l
47 Nominal Steam Generator Vass - 10% 20.9 441 504 1.87 48 Nominal RCS Flovi Rate + 10% 21.3 443 503 1.86 49 4.0 ft' Double ended Rupture 20.8 441 505 1.84 50 3.0 ft' Double ended Rupture 20.7 442 506 1.89 i 51 1.4 ft' Double ended Rupture 20.2 443 508 1.98 ,
52 1.0 ft' Double ended Rupture 17.4 453 511 2.38 ;
i 53 0.5 ft' Double ended Rupture 12.4 472 514 2.48 54 0.1 ft' DouNe ended Rupture 5.3 498 517 6.30 56 1.2 ft' Spht Break 18.9 448 510 2.06 57 1.0 ft' Spht Break 17.5 453 511 2.06 58 0.5 ft' Split Break 12.5 472 514 2 18 B
SS 0.111 ft' Spht Break 5.6 498 517 6.30 j 61 i Open Main Steam Safety Valve 5.5 498 517 5.34 l 62 MSIV Stroke Time = 10.0 seconds 22.8 439 498 1.97 63 Mam Feedwater isolation Valve Stroke Time = 10.0 seconds 21.2 440 503 1.88 64 MSIV and FWlV Stroke Times = 10.0 seconds 23.8 439 495 1.95 65 Delay Safety injection until Low Pressurizer Pressure - Si Setpoint 20.9 441 504 1.84 i e 66 Loss of Offsete Power Coincident with Safety injection 11.7 392 431 4.56 I i
l t
i 17 l
t
- , ..y, y,.r
Question 15:
What is the fuel gap conductance used in each of the steam line t ,k ccses analyzed? !
Response: l For the demonstration steamline break analyses presented in Reference A, a i Iow value of the gap conductance was used. During the at-power and zero :
power transients, the power increase is mitigated by the Doppler fuel 1 temperature reactivity feedback. For cooldown events, the use of high values !
of the gap conductance would result in a lower Doppler defect; thus, the peak power would be higher. This sensitivity has been demonstrated by varying the Doppler defect and through sensitivity studies performed for the " Excessive j increase in Steam Flow Event" analysis (presented in Reference D). Future
- MSLB analyses will use a conservatively high value of the gap conductance. !
I Question 16: !
Explain Figure 3.4.51 and identify differences between the case presented and !
the reference case. [
Response
The reference case is a 5.6 ft 2double-ended rupture (DER); the case presented ,
in Figure 3.4.51 is a 1.4 ft 2 DER. Due to the smaller break flow area, the amount of steam released from the " intact" steam generators is reduced for the :
1.4 ft2 DER prior to the closure of the main steam isolation valves. Following the closure of the MSIVs, only one steam generator continues to depressurize, and the flow rate is limited by the integral flow restrictor with the 1.4 ft 2 effective flow area; thus, the steam release rates for the two cases are similar.
l 18
Question 17:
Provide comparative figures of the RETRAN and Final Safety Analysis Report (FSAR) results. Similarly, provide a table showing the initial conditions and assumptions used in the RETRAN reference cases and FSAR cases. If there are any differences, discuss the reasons.
Response
The requested figures we provided as Figures 17-1 through 17-8. With the following qualifications, the assumptions used in the Reference Case, including the degree of reactor vessel inlet and outlet plena mixing and the reactivity weighting factors between the intact / faulted quadrants, and the assumptions used in the vendor's analysis are similar:
Not all details of the vendor's modeling methodology (e.g., how the plena mixing is actually modeled) are known to TU Electric, hence, there may be some differences induced by model differences.
The vendor's calculation of the delay time for boron to reach the core is not accurate. The vendor predicts the boron to reach the core after the power plateau is attained, but well before the TU Electric prediction.
The actual CPSES safety injection flow rates are less than assumed in the vendor's analysis. The vendor has provided evaluations to justify the adequacy of the current flow rates.
A low compensated steamline pressure setpoint of 450 psia was used in the vendor's analysis. Subsequent evaluations were provided to justify the value of 395 psia used in the TU Electric analyses.
l i l V 19 I
l
_ _ _ _ . _ . . ~ . _ _ _ . . - _ - . - . _ - _ . --- -- . - - - - - - - - - - - - - - -
J I
)
a
- I i
I l 0.25 !
c 0.2- -
- I o
n.
0.15-l 5? 0.1 - SAft i g .!E U 0.05-os 2C -
0-0 5'O 100 150 200 250 TIME (SECCFCS) !
Figure 17-1 r
0.25 .
~ 'l x 0.2- ,
3 5 d 0.15-92 )
6 15 0.1 - l 2e f1AR l E 0.0s-Oes :
Og 0
'l 1
O 50 100 150 200 2 51
- l TIME (SECCNDS) :
Figure 17-2 ,
I l
l 2500 l W
c: 2250-3 l
$ 2000- F5A' I w
e 1750-
- a. g ea w-1500-E:!
- 1250-m
@ 1000-
_l 0
m 7% . . . .
' 0 50 100 150 200 250 TIME (SECCNDS)
Figure 17-3 20
! .. .. l l
l l
4 i
l l
E i IIi 500- -
"f _ 400- -
cC yy 300-
/ PAR
.e bd E"
@ 100- g -
e -
1 0 , , , , .
0 60 100 150 200 250i TIME (SECCNDS)
Figure 17-4 l
l l l
3000 ,
j 2000- 1
! I
} g 1000-
! EE O
^
l UW l
@~
c
~1000~ f.$A g
.M-
-3000 . . , , , ,
1 0 50 100 150 200 250 TIME (SECCNDS) l Figure 17-5 l
.i l
l S00 .
g 550-
~
y 0#
500-
$ 'uJ
- ! E .tso- w e&J< .50 0- /
i t
Wx Ww FSAR I
350- _ .uract sactran E i
@ n.
> 3. 300-w 250 200 250 0 50 100 150 TIME (SECCNDS)
Figure-17-6 21
l e
l l
I l
l f 1100 l 1000 _gy yg y 9004 /SAA 3 800- .
$_ 700- [
W! S 500-4 EE 500- ggg 1~ 100-
$ 300- % .
5 200-100 , , , ,
0 50 100 150 200 250 T1ME (SECCNDS)
Figure 17-7 l 3.5 '
-"" '"E E e
o 3-l 3: 5 2.5-O5 a 2- F.SAR
- u. z 2 '5
<=
1.5- h/
w _o 1-ui U 0.5-S E 0 i
0 5'O 160 150 200 250 TIME (SECCNDS)
Figure 17-B l
l l I l
l 22 l
1
~
l i
t i
l References I
l A. " Methodology for Reactor Cora Response to Steamline Break Events," ;
RXE-91-005, TU Electric, May 1991.
B. " Steady State Reactor Physics Methodology," RXE-89-003-P I (Approved), TU Electric, July 1989.
C. "VIPRE-01 Core Thermal Hydraulic Analysis Methods for Comanche Peak Steam Electric Station Licensing Applications," RXE-98-002 (Approved),
I TU Electric, June 1989.
D. " Transient Analysis Methods for Comsnche Peak Steam Electric Station l Licensing Applications," RXE-91-001 (Approved), TU Electric, February, ;
1991.
i I
5 I
23
_ _ _